Monday, November 03, 2014

Pennsylvania Computer Nerd Claims Berkowitz Lost Because of 2008 Election Fraud

With the 2014 election climaxing tomorrow, it seems appropriate to look at John Foelster's claims about the 2008 election in Alaska.  Foelster left comments on some Alaska blogs, including this one, last week, with a link to his website  [UPDATE Nov 23, 2016: his website now has restricted entry]  where he spells out the data that underlies his assertions. He writes
The fraudulent vote count would have resulted in Don Young being wrongly seated in a House seat actually won by his Democratic opponent Ethan Berkowitz, and would certainly have changed the outcome of the District 7 Lower House Race so that Democrat Karl Kassel would have beaten Republican Mike Kelly. [Note Kassel lost by four votes.] Democrats Andrea Doll of the 4th lower House District and Val Baffone of the 28th Lower House District were also likely the actual winners of the races they officially lost.
At what point do we take assertions seriously?  After all, I had had emails about the National Guard scandal as far back as 2010 and I'd seen Blaylock's long list of allegations on line.  But I didn't write about it for lack of further information and lots of other things to do.

My experience with the 2012 Municipality of Anchorage election (see list of posts at bottom of this post) showed me how vulnerable the voting machines are.  While I tend to think the problems in that election were more related to incompetence and not tampering with the voting machine software, the situation exposed the many vulnerabilities of the procedures and the machines.

I've read Foelster's claims.  They represent a lot of painstaking work, not only in gathering and interpreting the data, but also in how to present it.  He has a series of video tapes that walk you step by step through his hypothesis and the evidence backing it up.  While the conclusion is sensational, his presentation is not.  It's painstakingly detailed, self aware and self effacing, and outlines exactly how he went about getting to his claims.

I've sent the information to a couple of good data people and got a long response from one who knows a lot about elections and the computers. While this person didn't read it all, and raised some questions here and there, he allowed it was a possibility and was concerned about the greater environment of election fraud vulnerability of the nation as a whole.  I've also had some email exchange with Foelster to follow up with questions I had.

Problems with Reporting Computer Crime

Manipulating computer data is the kind of crime that doesn't emotionally effect people like murder, armed robbery, kidnap, and rape.  Television footage of computer code just isn't compelling. Computer crime is all hidden in 1's and 0's inside the computer, in computer code that most people don't understand.  It's also something we don't want to believe is happening because it violates our sacred belief in American democracy.  There are lots of websites that give details on how the voting machines can be hacked.  Bradblog covers all sorts of voting issues including problems with voting machines.  Here are some others.
I list all these sites just to remind people that there are legitimate and serious issues with voting machines that should be understood by everyone.  Citizens United is one issue, but the vulnerability of the voting machines may be an even bigger one.

So, back to Foelster's allegations.  Here, from his website, is the outline of his evidence:
[UPDATE Nov 23, 2016: his website now has restricted entry]
"The lines of evidence are as follows:
  1. The polls in 2008 indicated that the Democrats in Alaska would do much better than they did in the reported results.
  2. In 2012, there was a large swing to President Obama, one that was too large to have been caused by Governor Palin no longer being on the Republican ticket.
  3. The number of Democrats in the electorate in 2012 was significantly lower than the number in 2008.
  4. The report on electorate composition this information can be found in appears to have been reformatted in an attempt to draw attention away from it.
  5. The 2008 results are unique in recent Alaska history for having very Democratic Absentee results and very Republican in precinct results.
  6. A variation on a known technique for compromising AV-OS machines would have produced the effects described above, and this hack could have been introduced by one person working in Juneau.
  7. The size of the above anomaly varies from district to district based on the number of registered Democrats in precincts with AV-OS voting machines.
  8. The State Review Board’s Hand Count Audit of the paper ballots could also have been compromised by this same single person."
Each item links to further information.  

As I read this, I see him taking two main approaches:
  1. Technical:  The data show anomalies that indicate votes in 2008 were tampered with.
    1. In 2012 there was a significant boost in votes in Alaska for Obama than in 2008.  This runs counter to every other state where the support for Obama went down in 2012. This happened because in 2008, a large number of Democratic votes in Alaska were switched over to Republicans. 
    2. How the voting machines work and how to hack them to get the results he thinks happened
  2. Human:  The narrative of who might have been involved, how, and why
For me, the human narrative part is weak.  But that doesn't really matter.  He doesn't have to prove who did this.  What he has to demonstrate is the problem with the data and technically how votes could have been manipulated.   And that part I think he's done - at least to the point that others with the appropriate expertise should follow up and determine if his allegations have merit.  It may lead to a dead end, but even then it would help expose the vulnerabilities of the process and technology further.

The technical part isn't necessarily flawless.  Perhaps other explanations would account for what he found.  For example, if enough Democrats switched parties to vote in the Republican primary race between Murkowski and Miller, would that accounts for the dip in Democrats that 2008?  I don't know. 

Who Is John Foelster?

Foelster identifies himself as a 'nerd' who lives in Pennsylvania and has never been to Alaska.  He also has Aspbergers. And experience in computers and voting technology. I don't see that he has any particular interest in Alaska politics other than he saw this inconsistency and then obsessively pursued it. My sense of Foelster is that this anomaly caught his attention and he ran with it.    I don't believe he works for any party or has any particular personal interest in Alaska.  It's just a puzzle that came his way and he got deeply into it.  I understand that.  I did something similar with the redistricting board.

Computer crimes are hard to prove - first that they happened and second who did them - without lots of access, patience, and savvy.  

Even if all his allegations proved to be correct, I don't see Ethan Berkowitz being retroactively sworn in to the House of Representatives.  But if this really did happen, we ought to know about it and take steps to do something to protect us better in future elections.  And if it didn't happen here the way Foelster says, there are other possibilities where it might have happened.

Can it happen in tomorrow's election?  Yes, but there were patches made, according to Foelster, in 2011 that would preclude the particular hack he describes as for 2008.

What Next?

I see two things that should be done here:
  1. We need a technical election committee that reviews every election in Alaska.  It looks at the hardware, the software, and makes a statistical analysis of voting results to find any suspicious anomalies.
  2. That committee could start by reviewing Foelster's work.

I would note that the Municipality had such a committee - though its charge was not quite this thorough - but it was abandoned before the 2012 election.

I would also note that in that election there were lots of sloppy practices - plastic seals on ballot bags that easily broke off, voting machines and ballots being taken home by election workers overnight - that opened up many opportunities for fraud.  Barb Jones, the new municipal clerk (after the 2012 election) and the election official Amanda Moser, have eliminated some of these practices and have been accessible to me as a blogger to exactly how each step of the process works.  But we're still too much in the small town, we trust each other, level that our procedures originally came from.  We're still leaving a lot of windows open and doors unlocked.

2012 Municipality Election Posts To Show Where My Concerns Come From

I have put together a list of posts I did on the 2012 Municipal election.  (The posts were not well labeled and I had to poke around to find them.)  I offer this list to help readers understand my experience with elections and why I'm willing to give John Foelster some attention here. 

The Myth of the Big Election Turnout 

Guadalupe Marroquin, Former Anchorage Election Chief Talks About The Election... (10 minute video on how to tamper with the machines and how they work from the previous election chief)

What Do The Election Percentages and Numbers Tell? Maybe Nothing

Polling Gap - Dittman Confirms It's the Biggest 

Brad Friedman Rips Apart Election Commissioner's Testimony 

Jacqueline Duke, Elections Chief, Fired by Assembly

What's Happening With The Anchorage Election? 

Assembly Exchanges Barbs: Barbara Jones To Replace Barb Gruenstein.

Citizen Group on Election Meeting Now with Assembly Attorney

Hensley Report on Election Now Available - Form Over Substance 

141 "Potentially Uncounted Ballots" Found July 11 (From April Election)

How Many Ways Are There To Steal An Election? And Why Doesn't Anyone Care?
(This is marginally about Anchorage - trying to link us to bigger national issues)


  1. Two quibbles.

    ALMOST all other states swung to Romney in 2012 besides Alaska. The ones Obama improved in were New York and New Jersey (probably because of Hurricane Sandy, along with Maryland, probably for the same reason and, incredibly, Louisiana and Mississippi, which is a tale in and of itself. (Briefly, the large and rapidly growing Black population in the Deep South and Obama's White support in these states already being at rock bottom in 2008.)

    The Assure 2.0 firmware patch installed in 2011 CLAIMS to have fixed the holes, but the patch was not, to the best of my knowledge, reviewed by people outside of DIEBOLD/Premier Elections. My guess is (and here things get frightfully technical) they fixed the negative integer overflow error allowing negative votes to be pre-loaded in the election counters but might have left the AccuBasic interpreter with the ability modify the counters contents. So we really don't know if the holes in the machines have been adequately fixed or not.

    As to the suggestion that registered Dems changed registration in 2010 to vote against Miller in the Primary, that's testable because there's a month by month account of how many people are registered in each party going back to 2003.

    The answer appears to be "No".

    Here's the last report from before the 2008 registration deadline.

    This one is Jamuary of 2010

    This one is August of 2010:

    And this is from the registration deadline from 2012:

    Over 2009 voter registration dropped by 6,941, with most parties losing people. The biggest gainer was the Libertarian Party of Alaska which shot from 6,972 to 9,280 registrants while Undeclared registration dropped by 7,397. The biggest loser in percentage terms would have been the Republican Moderates who dropped nearly a quarter of their number from 3,939 to 3,088. Democrats went down a net of 923 voters and Republicans went up 109. A lot of this probably had to do with the voter rolls purges that happen every March and people who die not being replaced with 18 year olds or new migrants who saw no reason to register in 2009.

    The net loss from January of 2010 to the beginning of August was 1,215 voters. This time the Democrats are the biggest loser, with a net loss of 1,004 people, followed by Undeclared dropping 971. This is not matched by an increase in Republicans though, they lost a net of 206. The two groups that had the largest gain were actually the Alaska Independence Party and the Non-Partisan, with 635 and 582.

    So if Democrats were switching to Republican and Non-Partisan for the Primary vote, they were more than offset by Republicans jumping ship to AI. (Presumably NOT to vote in the ADL Primary.)

    Between 2010 and 2012, the voting rolls surge up 18,859, 10,882 of which are new Republicans, AI posted impressive gains again, going up 1,229 voters while the Libertarians lost most of their gains from 2009-2010, Non-Partisan and Undeclared went up 4,673 and 8,852 persons as well, while the Republican Moderates disbanded entirely. Democrats also dropped 1,465 persons concurrently. This looks like a pretty serious voter registration drive by conservatives for the 2012 campaign unmatched by anything on the Democratic side..

    For the record, here's now:

    Net gain of 2,577 voters, with the churning being away from R and D and towards AI, Non-Partisan and Undeclared.

    Also, two persons have joined the Twelve Visions Party. Judging by its website, this national group is giving Vermont's VoteKISS party a run for its money as the trans-national affiliate of the Monty Python Very Silly Party.

  2. This may be a little off topic, but what about that "AD" on page A5 of the sunday Anchorage Dispatch telling Christian Conservatives to vote. There is no name on it, saying who paid for the ad. As far as I know, it was commentary from the editors of the paper.

    1. Since the ad is simply telling people to vote and not how to vote for any particular candidate or proposition, I suspect they don't have to say who paid for it. But I'm not certain.


Comments will be reviewed, not for content (except ads), but for style. Comments with personal insults, rambling tirades, and significant repetition will be deleted. Ads disguised as comments, unless closely related to the post and of value to readers (my call) will be deleted. Click here to learn to put links in your comment.